Wonder

Wonder is an upscaling image repair model designed for small to medium-sized low-resolution images that are soft, noisy, or affected by compression artifacts. It helps restore clean, realistic detail without the artifacts that can occur when stacking multiple enhancements.

Wonder v1 or v2 are not intended and are not supported for RAW files or images from modern cameras. For those files, use enhancements such as RAW Denoise, Denoise, Sharpen, Recover Faces, and Upscale. While Wonder 3 can be used on high-resolution or RAW images at the Low setting, the changes will generally be minimal.

The maximum output for this Wonder generative model is 128MP.

If you experience rendering issues, or are unsure whether Wonder is the right model for your image, contact Support and include the original file so we can recommend the best workflow and confirm which model is best suited.

How to access the model

You can access Wonder from the "Select a Filter" button in the right panel.

Note: if Auto-Pilot is on, this "Select a Filter" button will be named "Choose something else"

Results examples on this heavily compressed file:

Wonder Versions

Wonder (v1): Best for heavily compressed or very low-quality images that need strong cleanup. Produces more aggressive, stylized results and can over-process in some cases.
Wonder (v2): Best for low-resolution or challenging images where realism is the priority. Produces cleaner, more natural results than v1, but can be more conservative on very low-detail images.
Wonder (v3): Released in Topaz Photo v1.5.0 - Best for recovering detail across a wider range of low to mid-quality images. Improves on v2 with stronger recovery, more realistic textures, and better handling of false resolution with minimal user input.

- Low setting: Best for preserving the original input data. It stays most faithful to the source, though it may appear slightly softer at 4x.
- Medium setting: A balanced middle ground.
- High setting: Can be more ‘creative’ with detail reconstruction. It’s best suited for heavily degraded images or large upscale factors, such as 6x.

Key things to know

Wonder works best with Cloud Processing: Because of the computational resources required to run this model, Wonder will give best processing speeds on the Cloud.
Wonder is a Single-step model with no preview controls
Not recommended for low-light phone images with heavy noise or already sharp and high-quality images.
Wonder cannot be used as a Preset, but it can be used in Batch Processing workflows. See details here.

Important: The Wonder models are for small or medium low-resolution files that need strong cleanup and enhancement and that have compression artifacts to fix. It is not to be used for high-resolution images, or for images shot on modern cameras, and is not for images that are already sharp. While Wonder 3 can be used on high-resolution or RAW images at the Low setting, the changes will generally be minimal.

Local Rendering

Wonder v1
Local rendering is available if you:

Have a Topaz Pro license or Founder status
Use a machine that meets the minimum system requirements

When these conditions are met, Wonder v1 can run locally on your machine.

Wonder v2 and v3
Local Wonder v2 and v3 require:

Windows: NVIDIA or AMD Graphic card with 8GB VRAM or more
Mac: Apple Silicon (M-series) machines with 16GB RAM minimum

Not supported on:

Intel Macs
Windows ARM / Snapdragon systems

Cloud Rendering

You can run Wonder (all versions) using Cloud Processing.

Provides faster processing then using local processing

Contact Support

If you have issues using the model, contact us at Support at help@topazlabs.com and we can troubleshoot. When you write in, make sure to provide:

The image you are trying to upscale. Make sure it's a small or medium low-resolution or compressed file, which will be a case for the Wonder model.
Details on the issue: Are you using Local or Cloud processing? Which version? Attach the file used to the email and send Details on the behavior and screenshots of the error you get, if any.
Your app's logs: Help menu > Open Logs folder > Attach all the logs you see in the email to Support

Upscale & Resize

Generative Models